Much has be written about the explosive growth in capital spending on AI by Google, Meta, Microsoft, and Amazon. (See: https://www.wsj.com/economy/the-ai-booms-hidden-risk-to-the-economy-731b00d6?mod=author_content_page_1_pos_1 , paywall) Those four companies are on tap to spend about $360 billion this year on building out data centers and related investments. Given the high prices of their stocks, market participants are assuming that AI investment will have a huge payoff in the future and are willing to overlook short term declines in free cash flow.
My sense is that investors are underappreciating three risks. The first is that the return on tangible capital will likely be lower than the hitherto high returns on intangible capital earned by the four companies. Second, the bidding war for talent where $100 million payouts have taken place will squeeze margins. Third, and perhaps most important, is that AI will not, at least initially, be a winner take all market. It is likely that AI will be nowhere near as profitable as winner take all Microsoft's operating systems, Google's search, and Meta's social media.
Why? At the present time there are six highly capitalized entrants in the AI market and many smaller ventures competing for market share. The six are ChatGPT, Gemini, Anthropic, Perplexity, Meta AI, and Grok. Although ChatGPT has the current lead, this is a highly fluid market environment with deep pockets competing for market share. Thus, my guess is that investors will soon be disappointed with the returns coming from AI. The likely beneficiaries of this competition will be consumers and the highly talented employees.
No comments:
Post a Comment